Individualizing Revascularization Strategy for Diabetic Patients With Multivessel Coronary Disease

BACKGROUND:In patients with diabetes and multivessel coronary artery disease (CAD), the FREEDOM (Future Revascularization Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease) trial demonstrated that, on average, coronary artery bypass grafting (CABG) was superior to percutaneous coronary intervention (PCI) for major acute cardiovascular events (MACE) and angina reduction. Nonetheless, multivessel PCI remains a common revascularization strategy in the real world. OBJECTIVES:To translate the results of FREEDOM to individual patients in clinical practice, risk models of the heterogeneity of treatment benefit were built. METHODS:Using patient-level data from 1,900 FREEDOM patients, the authors developed models to predict 5-year MACE (all-cause mortality, nonfatal myocardial infarction, and nonfatal stroke) and 1-year angina after CABG and PCI using baseline covariates and treatment interactions. Parsimonious models were created to support clinical use. The models were internally validated using bootstrap resampling, and the MACE model was externally validated in a large real-world registry. RESULTS:The 5-year MACE occurred in 346 (18.2%) patients, and 310 (16.3%) had angina at 1 year. The MACE model included 8 variables and treatment interactions with smoking status (c = 0.67). External validation in stable CAD (c = 0.65) and ACS (c = 0.68) demonstrated comparable performance. The 6-variable angina model included a treatment interaction with SYNTAX score (c = 0.67). PCI was never superior to CABG, and CABG was superior to PCI for MACE in 54.5% of patients and in 100% of patients with history of smoking. CONCLUSIONS:To help disseminate the results of FREEDOM, the authors created a personalized risk prediction tool for patients with diabetes and multivessel CAD that could be used in shared decision-making for CABG versus PCI by estimating each patient's personal outcomes with both treatments.